Token导航 LogoToken导航

Agent Skill · 研究检索

design-sprint(设计 sprint)

用于处理 Jira 项目、任务、缺陷、Sprint、负责人和状态流转。它适合让 Agent 辅助查询工单、汇总迭代进展、创建任务或整理需求和缺陷信息。使用时要确认项目权限、字段配置和工作流规则,不同团队的 Issue 类型、状态和必填字段可能不同;涉及批量改状态、改负责人或创建工单时,应先预览变更内容再执行。

design-sprint 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 4,504。

总安装 31,824周安装 1,324GitHub Stars 779下载 10,296核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:design-sprint(设计 sprint)
来源仓库:https://github.com/wondelai/skills
仓库路径:skills/design-sprint
安装命令:
npx skills add https://github.com/wondelai/skills --skill design-sprint
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/wondelai/skills --skill design-sprint

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

五天压缩版设计冲刺框架,用于快速验证商业假设并通过原型测试用户反馈。适合高风险产品决策前的快速实验,避免数月开发后才发现方向错误。集成地图绘制、草图、故事板与测试环节,形成高强度聚焦流程。安装需通过 npx 添加指定 GitHub 仓库,适用于创业公司与创新项目。

design-sprint 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/design-sprint

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/wondelai/skills --skill design-sprint;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

design-sprint 怎么安装?

在终端执行 npx skills add https://github.com/wondelai/skills --skill design-sprint。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

design-sprint 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

design-sprint 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

design-sprint 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

design-sprint 用的人多吗?

公开数据是累计安装 31,824 次,近一周安装 1,324 次,下载 10,296 次,GitHub 779 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

design-sprint 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
design-sprint

设计 sprint · 本页

31,824779只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Design Sprint Framework

A five-day process for answering critical business questions through design, prototyping, and testing ideas with customers. Developed at Google Ventures and used by Google, Slack, Airbnb, and hundreds of startups.

Core Principle

Great solutions require both deep work and fast iteration. The Design Sprint compresses months of debate, design, and testing into a single week, creating focus and urgency that eliminates endless discussion.

The foundation: Traditional product development wastes months building the wrong thing. Design Sprints de-risk product decisions by testing with real users before writing production code.

Scoring

Goal: 10/10. When planning or executing a Design Sprint, rate it 0-10 based on adherence to the principles below. A 10/10 means proper structure, time-boxing, prototyping, and user testing; lower scores indicate skipping steps or insufficient testing. Always provide the current score and specific improvements needed to reach 10/10.

The 5-Day Sprint Process

Monday → Tuesday → Wednesday → Thursday → Friday
  Map      Sketch     Decide      Prototype    Test

Prerequisites:

  • Big challenge: Important problem worth a week's focus
  • Right team: Decision maker + 4-7 people with diverse expertise
  • Time commitment: 5 full days (10am-5pm), no interruptions
  • Space: Dedicated room with whiteboards

Sprint Master: One person facilitates, keeps time, manages energy.

Monday: Map

Goal: Understand the problem and choose a target for the week.

Morning: Start at the End

Exercise: Long-term goal

  • Write the sprint question: "What do we want to be true in 2 years?"
  • Example: "Customers use our product daily" or "We've captured 20% market share"

Exercise: Sprint questions

  • List obstacles and unknowns as questions
  • Example: "Will customers trust us with payment info?" or "Can first-time users figure out the interface?"

Format: Write on whiteboard, entire team contributes

Afternoon: Map the Challenge

Exercise: Map the customer journey

  1. List actors (different types of customers/users)
  2. Draw the journey from start to finish (left to right on whiteboard)
  3. Keep it simple: 5-15 steps max
  4. Example: "Hears about product → Visits site → Signs up → First use → Becomes regular user"

Exercise: Ask the Experts

  • Interview team members with specialized knowledge
  • CEO, designer, engineer, customer support, sales
  • Take detailed notes on whiteboard
  • Capture "How Might We" notes (HMW)

Exercise: How Might We (HMW) notes

  • Rephrase problems as opportunities
  • "Customers don't understand pricing" → HMW make pricing immediately clear?
  • Write each HMW on a sticky note
  • Vote on best HMWs, organize on map

End of Day: Pick a Target

Exercise: Choose the target

  • Which part of the map (customer journey) will you focus on?
  • Where's the biggest risk or opportunity?
  • Example: "We'll focus on the first 10 minutes after signup"

Decider: The person with authority makes the final call.

Monday output:

  • Long-term goal
  • Sprint questions
  • Customer journey map
  • Expert insights
  • HMW notes organized
  • Target customer and moment

See: references/monday.md for detailed Monday exercises and facilitation.

Tuesday: Sketch

Goal: Generate solutions. Each person sketches a detailed solution.

Morning: Lightning Demos

Exercise: Find inspiration

  • Look at competitors and analogous products
  • 3-minute demos: "Here's what I found, here's why it's interesting"
  • Capture good ideas on whiteboard
  • Don't limit to your industry—borrow from anywhere

Exercise: Divide or swarm

  • Divide: If map has multiple parts, different people tackle different sections
  • Swarm: If one critical problem, everyone tackles the same thing
  • Most sprints = swarm

Afternoon: The Four-Step Sketch

Goal: Everyone individually sketches a detailed solution (not as a group!)

Step 1: Notes (20 minutes)

  • Walk around room, review map, HMWs, inspiration
  • Take notes silently

Step 2: Ideas (20 minutes)

  • Rough doodles, mind maps, stick figures
  • Quantity over quality
  • Still working alone

Step 3: Crazy 8s (8 minutes)

  • Fold paper into 8 sections
  • Sketch 8 variations in 8 minutes (1 minute each)
  • Forces you past first idea
  • Can be 8 variations on one idea or 8 different ideas

Step 4: Solution Sketch (30-90 minutes)

  • 3-panel storyboard showing customer experience
  • Step 1 → Step 2 → Step 3 (beginning, middle, end)
  • Make it self-explanatory (someone should understand without you explaining)
  • Use text, arrows, simple drawings
  • Give it a catchy title
  • Anonymous: Don't put your name on it

Critical: No group brainstorming. Individual work produces better, more diverse ideas.

Tuesday output:

  • Each person has a detailed solution sketch
  • Sketches are anonymous and self-explanatory

See: references/tuesday.md for sketching templates and examples.

Wednesday: Decide

Goal: Critique solutions and choose the best one to prototype and test.

Morning: Sticky Decision

Exercise: Art museum

  • Tape solution sketches to wall
  • Give everyone dot stickers
  • Silently review sketches (no talking!)
  • Put dots next to interesting parts

Exercise: Heat map review

  • Discuss each sketch for 3 minutes
  • Facilitator narrates: "Here they see X, then click Y..."
  • Sketcher stays silent (don't reveal yourself yet)
  • Team calls out interesting parts
  • Scribe captures standout ideas on whiteboard

Exercise: Straw poll

  • Each person votes for one solution (put one large dot)
  • Explain your vote in 1 sentence
  • This is non-binding, just to see preferences

Decider: Person with authority gets three large dots (supervote). Their decision wins.

Afternoon: Rumble or All-in-One

If multiple winners:

  • Rumble: Competing prototypes (test different approaches)
  • All-in-One: Combine best ideas into one prototype

Most sprints: All-in-one (simpler to prototype and test)

Exercise: Storyboard

  • Draw 10-15 panel storyboard (comic book style)
  • Each panel = one screen or step
  • Opening scene: How customer discovers you
  • Middle: Your solution in action
  • Ending: Successful outcome
  • Include just enough detail for Friday's prototype

Storyboard rules:

  • Keep it simple
  • Use stick figures
  • Words and arrows okay
  • Get specific about UI
  • 10-15 panels max

Wednesday output:

  • Winning solution(s) chosen
  • Detailed storyboard ready to prototype

See: references/wednesday.md for decision exercises and storyboard templates.

Thursday: Prototype

Goal: Build a realistic facade. You need something to test on Friday.

Prototype mindset:

  • Fake it
  • Prototype only what you'll test
  • Goldilocks quality: not too high, not too low (realistic enough to get honest reactions)
  • One day only

Prototype fidelity:

  • Too low: Sketches, wireframes (customers can't react realistically)
  • Too high: Working code, pixel-perfect design (wastes time)
  • Just right: Looks real, doesn't work real (facades, click-through, video)

Assign Roles

Makers (2+ people):

  • Designer, writer, asset collector (images, icons)
  • Build the prototype

Stitcher (1 person):

  • Combines pieces into final prototype
  • Usually in Keynote, Figma, or prototyping tool

Writer (1 person):

  • Writes all copy
  • Headlines, button labels, descriptions

Collector (1-2 people):

  • Gathers assets (photos, icons, competitor screenshots)
  • Provides raw materials

Interviewer (1 person):

  • Writes interview script for Friday
  • Practices interviewing

Sprint Master:

  • Helps where needed
  • Keeps energy up

Build the Prototype

Tools:

  • Web/App: Figma, Keynote, PowerPoint (linked slides)
  • Physical Product: Video walkthrough, 3D-printed mockup
  • Service: Role-play video, scripted interaction

Thursday morning:

  • Divide storyboard into scenes
  • Assign scenes to makers
  • Start building

Thursday afternoon:

  • Stitch together
  • Review as team (does it match storyboard?)
  • Rehearse for Friday (run through entire flow)
  • Trial run (test with someone not on sprint team)

Prototype checklist:

  • Follows storyboard exactly
  • Looks real enough to get honest reactions
  • Can walk through in 5-15 minutes
  • Interviewer knows how to present it
  • Trial run completed

Thursday output:

  • Realistic prototype ready to test
  • Interview script written
  • Interview room prepared

See: references/thursday.md for prototyping tools and techniques.

Friday: Test

Goal: Interview 5 customers, learn what works and what doesn't.

Setup

Interview room:

  • Quiet space with table, 2 chairs
  • Laptop with prototype
  • Camera recording screen and customer face

Observation room:

  • Separate room with live video feed
  • Team watches together
  • Whiteboard for notes

Roles:

  • Interviewer: Conducts all 5 interviews
  • Team: Watches, takes notes

The Five-Act Interview

Act 1: Friendly Welcome (5 min)

  • Greet warmly
  • Explain you're testing prototype, not them
  • Ask permission to record
  • Encourage thinking aloud

Act 2: Context Questions (5 min)

  • Ask about their background
  • Example: "Tell me about how you currently handle [problem]"
  • Goal: Understand their mindset and current behavior

Act 3: Introduce the Prototype (5 min)

  • Show landing page or entry point
  • "What's this? What do you think it's for?"
  • Don't explain—let them interpret
  • Note: Do they get it?

Act 4: Tasks and Nudges (15 min)

  • Give open-ended task: "Go ahead and explore"
  • Follow with specific tasks from storyboard: "Try to [complete action]"
  • Use nudges when stuck: "What would you do next?" or "What's going through your mind?"
  • Don't help—watch them struggle
  • Encourage thinking aloud

Act 5: Debrief (5 min)

  • "What did you think overall?"
  • "Who is this for?"
  • "What worked? What was confusing?"
  • Ask about specific parts you're uncertain about

Interview length: ~30 minutes per customer

Between interviews:

  • 30-minute break
  • Team discusses observations
  • Update questions if needed

Five Is the Magic Number

Why 5 customers?

  • Patterns emerge after 3-5 people
  • Diminishing returns after 5
  • Doable in one day (5 × 1 hour = 5 hours with breaks)

Who to recruit:

  • Target customers (match your personas)
  • Screener survey to qualify
  • Incentive ($100-$200 for B2B, $50-$100 for B2C)
  • Schedule 6 (expect 1 no-show)

Take Notes: Pattern Recognition

While watching interviews, team captures:

Column 1Column 2Column 3Column 4Column 5
Customer 1 notesCustomer 2 notesCustomer 3 notesCustomer 4 notesCustomer 5 notes

Mark with ✓, ✗, or ~:

  • ✓ Positive reaction, success
  • ✗ Negative reaction, failure
  • ~ Neutral or mixed

After all 5 interviews:

  • Look for patterns (did all 5 struggle with the same thing?)
  • Count ✓ ✗ ~ per row
  • Identify what worked and what failed

End-of-Sprint Debrief

Organize findings:

✓ What worked:

  • Features/flows that all customers understood
  • Messaging that resonated
  • Design that felt intuitive

✗ What failed:

  • Confusing terminology
  • Missing steps
  • Wrong assumptions

~ Mixed results:

  • Some got it, some didn't
  • Unclear if it matters

Next steps:

  • If core concept validated: Build it (or next sprint on details)
  • If major issues: Pivot or next sprint to solve problems
  • If totally failed: Back to drawing board (but you saved months!)

Friday output:

  • Interview videos
  • Pattern notes
  • Clear list of what works, what doesn't
  • Decision on next steps

See: references/friday.md for interview scripts and note-taking templates.

When to Run a Design Sprint

Run a sprint when:

  • High-stakes decision
  • Not enough time to build and test normally
  • Team is stuck in endless debate
  • Multiple solutions possible
  • New product, feature, or major redesign
  • Need to de-risk before investing

Don't run a sprint when:

  • Problem is clear and solution is obvious
  • You just need to execute
  • Team isn't bought in
  • Can't get decision maker for full week

Variations

4-Day Sprint:

  • Day 1: Map + Sketch (compressed)
  • Day 2: Decide
  • Day 3: Prototype
  • Day 4: Test

Remote Sprint:

  • Use Miro/FigJam for whiteboarding
  • Zoom for meetings
  • Same schedule, digital tools

Multi-Sprint:

  • Sprint 1: Broad problem, choose direction
  • Sprint 2: Deep dive on chosen solution
  • Sprint 3: Refine details

Common Mistakes

MistakeWhy It FailsFix
Skip prototypingNothing to testAlways prototype, even if simple
Over-engineer prototypeWaste time on details that don't matterFacade only, not working code
Test with wrong usersInvalid feedbackScreen for target customers
Explain prototype to usersDefeats the testLet them struggle, observe confusion
No decision makerCan't commit to decisionGet Decider for full week or don't sprint
InterruptionsBreaks focusProtect the week, no meetings/emails

Quick Diagnostic

Audit any sprint plan:

QuestionIf NoAction
Do we have a Decider for full week?Sprint will failGet commitment or postpone
Is the problem important enough?Waste of timeOnly sprint on big challenges
Can we prototype in 1 day?Wrong problem for sprintChoose more concrete problem
Can we recruit 5 target users?Can't test properlyStart recruiting now (2 weeks ahead)
Will team commit to no interruptions?Won't maintain focusGet buy-in from leadership

Reference Files

Further Reading

This skill is based on the Design Sprint process developed at Google Ventures. For the complete methodology, exercises, and case studies:

About the Author

Jake Knapp created the Design Sprint process while at Google, where he ran sprints on products like Gmail, Chrome, and Google X. As a design partner at Google Ventures (now GV), he refined the process by running over 100 sprints with startups in the GV portfolio. The Design Sprint is now used by teams at Google, Slack, Airbnb, LEGO, and thousands of companies worldwide. Jake is also the author of *Make Time*, a framework for focus and energy.

平台分布

Codex

33.05%
按下载量换算3,403

Claude

28.45%
按下载量换算2,929

Cursor

19.65%
按下载量换算2,023

Gemini CLI

8.87%
按下载量换算913

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills